# Autka Data Sources and Marketplace Integrations

> Understand Autka's lawful data boundary, partner feed options, deep-link fallback, and checklist for enabling a live vehicle source.

Autka only aggregates vehicle offers when the project has a lawful, reliable right to ingest and display them. Scraping marketplaces or reverse-engineering private endpoints is deliberately outside the project boundary.

## Two separate paths

1. **Catalogue data** comes from licensed, partner-provided, seller-authorised, or otherwise permitted feeds. The backend normalizes those records into Autka's shared offer model.
2. **Marketplace deep-links** stay on the client. Search filters are translated into the marketplace's own search URL, and Autka opens that site without ingesting listing data.

<Note>
  Production builds do not fabricate inventory. The mock source is for local development and debug/demo use only; unsupported marketplaces remain available through verified deep-links.
</Note>

## Partner integration options

Autka can work with dealer/importer JSON or CSV snapshots, licensed pull APIs, signed partner pushes, connected seller accounts, and broker or logistics feeds for the USA → Poland workflow.

A live catalogue source should provide stable source and listing IDs, title, make/model, price and currency, fuel, transmission, region, original listing URL, plus explicit update and expiry semantics. Year, mileage, location, coordinates and images improve the result when available.

## Acceptance checklist
Before a real source is enabled, record at least:

- the right to ingest and display the data;
- caching and image/text reuse rules;
- attribution, branding, and link-back requirements;
- stable IDs plus deletion/expiry semantics;
- full-snapshot versus delta behavior;
- pagination, rate limits, retries, and expected freshness;
- currency, tax, and unit conventions;
- lead, affiliate, or reporting requirements;
- production credentials stored only as backend secrets.

If those answers are unclear, the source stays disabled and Autka uses a deep-link instead.

## Moving a source live

When a lawful feed becomes available, document its contract, add and test a backend ingestion adapter, declare cleanup semantics, normalize records into the shared offer model, keep credentials server-side, and verify source health before enabling production ingestion.

The maintained implementation details remain in the Autka repository's `docs/DATA_SOURCES.md` and backend documentation; this page summarizes the product-facing boundary.